Trace2Skill: 将轨迹局部经验转化为可迁移的代理技能
Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills
AI总结：
本文提出Trace2Skill框架，通过归纳推理将广泛执行轨迹整合为统一的技能目录，有效提升代理技能的可迁移性和实用性，适用于多种领域。
摘要：
大型语言模型（LLM）代理日益依赖领域特定技能，但手动编写此类技能难以扩展，而纯参数知识生成的技能常忽略关键操作陷阱。我们引入Trace2Skill框架，通过归纳推理将广泛执行轨迹整合为统一的技能目录。Trace2Skill支持深入现有人工编写技能和从弱LLM生成草稿中创建有用技能。实验表明，Trace2Skill在多样化的领域中均表现出色，包括办公流程、数学推理和视觉问答。重要的是，进化出的技能不仅限于所用轨迹的简单记忆：它们在不同模型规模、不同模型家族和非分布设置中均能迁移。例如，从Qwen3.5-35B轨迹进化出的技能使Qwen3.5-122B代理在WikiTableQuestions任务上提升高达57.65个百分点。进一步分析显示，Trace2Skill优于序列技能编辑和ReasoningBank式检索记忆，能将重复失败和 workaround 压缩为标准操作程序（SoPs），并产生可重用的技能，无需参数更新或测试时检索。
英文摘要：
Large Language Model (LLM) agents increasingly rely on domain-specific skills, yet manually authoring such skills does not scale, and skills generated purely from parametric knowledge often miss critical operational pitfalls. We introduce Trace2Skill, a framework that consolidates broad execution trajectories in parallel into a unified skill directory through inductive reasoning over agent experience. Trace2Skill supports both deepening existing human-written skills and creating useful skills from weak LLM-generated drafts. Experiments demonstrate the effectiveness of Trace2Skill across diverse domains, including office workflows, math reasoning, and vision QA. Importantly, the evolved skills are not merely memorized artifacts of the trajectories used to create them: they often transfer across model scales, across model families, and to out-of-distribution settings. For example, skills evolved from Qwen3.5-35B trajectories improve a Qwen3.5-122B agent by up to $57.65$ percentage points on WikiTableQuestions. Further analyses show that Trace2Skill outperforms sequential skill editing and ReasoningBank-style retrieval memories, compresses recurring failures and workarounds into standard operating procedures (SoPs), and yields portable skills that can be reused without parameter updates or test-time retrieval.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
